In this paper, we have shown the development of a machine translation (MT) evaluation metric using quantum machine learning. For creating the training dataset, we have used the standard corpus and registered outputs of some MT engines across language pairs. Overall 85,600 translations were captured across 18 language pairs. These translations were then evaluated by human annotators and by some automatic evaluation metrics. The results of the metrics became the feature vector for the quantum regression model which was trained and the averaged human evaluation score became the target value. For training the regressor, variational quantum regressor was used at it is proved to be more efficient with NISQs. Two versions of the regressor were trained using two different optimizers, viz., COBYLA and ADAM. The results of the trained models were compared with chrf++ and COMET and were found to be at par with these standard MT evaluation metrics.